Expert: The new power system calls for vertical-type large models
2026-09-07 08:00Models🔥 42.2 heat score
1sources
1days unfolding
42.2heat score
4mentions
SummaryAI generated
On September 4, 2026, Yang Ping from South China University of Technology introduced the new energy AI vertical large model required for a new power system at the Taiyuan Energy Low-Carbon Development Forum. To address the issue that general large models struggle to adapt to physical constraints in electricity generation, the university proposed a hybrid architecture large model that integrates communication, artificial intelligence control, and electrical knowledge, which is unique worldwide. This model requires training and verification using high-quality datasets and a dedicated full-link production platform. Currently, with the support of departments such as the Shanxi Provincial Energy Bureau, the team has established the country’s first city-level integrated source-grid-load-storage platform in Changzhi, achieving high-precision data collection, analysis, and optimized scheduling, providing practical experience for the implementation of this model.
On September 4, Yang Ping from South China University of Technology introduced the new energy AI large model required for the new power system at the Digital Intelligence Development Forum of the Taiyuan Energy and Low-Carbon Development Forum. To address the issue that traditional general-purpose large models struggle to adapt to the physical constraints of the new power system, the university proposed a hybrid architecture large model that integrates communication, artificial intelligence control, and power knowledge for the first time globally. This model requires high-quality datasets and a fully integrated production platform to support training and verification. Currently, with the support of departments such as the Shanxi Provincial Energy Bureau, the team has established the country’s first urban-level integrated platform for energy, grid, load, and storage in Changzhi, achieving high-precision data collection, analysis, and optimized scheduling, providing practical experience for the implementation of this model.